Trimming ganged potentiometers with... trimpots?
It seems the most common usage of ganged potentiometers is for multichannel volume control, in which case you want a 0db attenuation option. However, suppose you want to control a variable resistance value in several channels at once, say for a filter applied to multiple channels, and the resistance range is >>0 to <infinity.
To bring for example a 0-50k pot into the correct range might require a 50k series resistor with each pot in the gang. Ganged pots are notoriously bad for matching between channels, but with 50k series resistance for each pot would it not make sense to use a PCB mount trimpot in series as well to obtain better matching? Such as a 40k fixed resistor and 10k trimpot? Or perhaps leaving the fixed resistor out completely? If there was concern about the coarseness of adjustment I suppose multiturn pots wouldn't break the bank.
I'd rather not get into "sound of different pots" discussions for now, and focus instead on practical problems. I can see two different methods for calibrating each channel; (1) measure the end resistance and trim so that all channels are matched, or (2) turn the ganged pot to a "median" or "most common" position (i.e., if the pot dial goes from 1-10, do the matching at 5), which should be a known value, and trim all channels to match that known value for that position. What pros/cons do you see for those two methods?
I'm not sure about the type of mismatching most common in pots. One would be a variation in end resistance with matched beginning resistances. i.e., three pots might read 0-50, 0-48, and 0-55. In that case each degree of turn would produce a different resistance change in each pot, and the trimming method above would simply shift the errors from one end to the other (or center the error, which in itself might be a good thing). The other would be shift in beginning resistance with good tracking between pots. i.e., three pots might read 2-50, 0-48, and 7-55. I somehow doubt things are that pretty.
Which error is most common? If it is the first, is the accumulation of error over the wiper range pretty linear, or is most of the variance in for example the first 10 degrees of wiper rotation?
Comments are most welcomed.
|
16th February 2005, 12:17 AM
Similar Threads
